CLI retry failed request and don't cancel entire flow on error

[zackpollard]

trafficstars

Currently the CLI will error and cancel the whole process if one request fails. We should probably instead retry a couple of times and then log an error that the asset failed, but continue on to other assets.

[masterT]

This is particularly annoying when uploading a lot of assets. Was there any progress toward this issue?

[masterT]

A significant improvement could be using an in-memory queue system (e.g., fastq) and having the ability to retry not just failed asset uploads but also other parts of the CLI.

I'm trying to upload an extensive library, and I have all kinds of network errors (I'm hosted with PikaPod, and my instance sometimes restarts, which makes the whole process crash).

I think I will craft a draft for this and open a PR.